A battery management system (BMS) is crucial for energy storage devices, improving battery packs’ efficiency and safer operation. Battery state of charge (SOC) and capacity are essential management factors in the BMS, and the estimation accuracy of SOC and capacity directly affects the regular use of energy storage devices during battery usage. This paper proposes the particle swarm optimization based multi-time scale dual extended Kalman filter (PMAEKF) for jointly estimating battery SOC and capacity. Multi-time scale dual extended Kalman filter (MAEKF) uses the information on micro and macro time scales to improve the accuracy and stability of the algorithm’s estimation. PMAEKF uses particle swarm optimization (PSO) to optimize the covariance matrix of the parameter transfer process across scales to improve the estimation accuracy of the algorithm and adapts to the difference between the measured voltage and the predicted voltage obtained by adaptive extended Kalman filter (AEKF) calculation. The difference between the measured and predicted voltage calculated by the AEKF is analyzed as a fitness function. Finally, the experiments under the Urban Dynamometer Driving Schedule (UDDS) proved that the proposed algorithm has better tracking and prediction ability in both SOC and capacity estimation.